CN101816007A - System for checking animation data, method for creating animation database, and system and program for registering animation data in animation database - Google Patents

Abstract

Provided is a collation system capable of effectively checking browsing movie data set in a site on a network in a browsable state. A check system for checking the read animation data set in a readable state in a website on a network, which comprises an animation database for registering a plurality of registered animation data as information; and a collating unit for collating the viewing moving image data with the registered moving image data registered in the moving image database, wherein the viewing moving image data has a still image displayed as a sample, the moving image database has a recording medium, and a registered image group composed of a plurality of registered images generated by a plurality of frame images of a reproduced moving image obtained by reproducing the registered moving image data is recorded as each registered moving image data, and the collating unit collates a retrieval image generated by the sample still image of the viewing moving image data with the registered image recorded in the recording medium of the moving image database.

Description

Animation data checking system, animation database creation method, registration system and program for registering animation data in animation database

technical field

The present invention relates to a collating system for collating browsing animation data set to be available for browsing at a site on the network, a method for creating an animation database in which a plurality of animation data used for checking browsing animation data is registered as information, And a registration system and program for registering animation data in an animation database.

Background technique

In recent years, with the digitization of information such as novels, music, movies, and photos, and the spread of the Internet, copyright infringement on the Internet has been regarded as a problem. The reason for this is that by digitizing information, it can be reproduced accurately without degrading the quality, and the secondary or tertiary use of information such as editing and adaptation is easy, and it can be distributed to the public via the Internet, making it possible to easily use other people's works. environment of. In particular, not only groups such as companies but also individuals are in such an environment, making prevention and management of copyright infringement difficult.

Recently, an animation sharing service for sharing animations has been provided on the Internet, and copyright infringement has become a big problem in the animation sharing service. The so-called video sharing service is a service that allows other users to download and share video data uploaded by a certain user, so that the video can be published and viewed. In this video sharing service, tens of thousands of pieces of video data are posted a day, and many of these video data include illegal video data that infringes copyrights. For example, uploading movies, TV programs, life images, promotional videos, etc. without the permission of the copyright owner.

At present, the management of copyright in animation sharing services is only about prohibiting the posting of copyright-infringing animation data content in the usage agreement, etc., and relies on the morality of each user, without setting up a dedicated inspection system. If even when illegal video data is uploaded, only the video data identified as illegal by a third party will be deleted. Copyright holders, etc. download and reproduce uploaded video data, and search and report illegal video data through viewing recognition. However, it is not realistic to check all video data that increases by tens of thousands every day. Furthermore, even if the illegal animation data is notified and deleted, users will often post it again, and existing countermeasures are ineffective.

However, most of the conventional video retrieval techniques are techniques in which a keyword is set in a video, a searcher inputs the keyword, and a video matching the keyword is searched for. However, in the video search method using a keyword, if the set keyword is inappropriate for a video, there is a problem that a suitable video cannot be searched. In addition, in the setting of keywords, there is also the problem that there are differences between individuals, and the keywords may not necessarily match due to differences in the subjectivity or expression of each person, and when the keywords are the same, there may be problems other than the target animation being retrieved. animation problem.

With respect to such a video retrieval method using keywords, the following video retrieval method is disclosed in Patent Document 1, which extracts a frame image of a switching video scene as a representative frame image, and calculates the value of the image from the whole or part of the representative frame image. The area of a specific color, the length of time between representative frames, the integral value of the sound intensity, the length of time that the sound is above a certain intensity, etc., assign codes corresponding to the characteristic values, and assign names to representative frames by code or code column, For the animation, use the name column of the frame corresponding to the representative frame column to make the name of the animation, and use the name of the animation for checking.

Patent Document 1: Japanese Unexamined Patent Publication No. 7-114567

Contents of the invention

However, at present, filtering of illegal video data is not sufficiently realized in video sharing services, and an effective video retrieval method is being sought. The animation retrieval method of the above-mentioned Patent Document 1 is aimed at retrieving consistent video scenes in an animation, and is not searching for illegal animation data in an animation sharing service.

In order to retrieve illegal video data in the video sharing service, it is necessary to investigate whether the video uploaded in the video sharing service is illegal, and for a plurality of videos uploaded in the video sharing service, it is necessary to check whether each video is illegal. As an uploaded animation, sometimes the original entire animation, sometimes only a portion of an arbitrary cut is uploaded. In addition, sometimes the same animation is uploaded multiple times, and continuous monitoring is required.

Patent Document 1 does not disclose a method of searching illegal video data in such a video sharing service. Moreover, in the animation retrieval method of Patent Document 1, since the animation is checked by the frame name column (animation name) corresponding to the representative frame column, it is necessary to download the uploaded animation and calculate the animation name of the uploaded animation. Check it out. Therefore, since the uploaded video must be downloaded every time, it takes time to search, and the download time becomes one of the bottlenecks for speeding up. Also, it may not be possible to correctly collate the animation data for small cuts, or it may not be possible to correctly collate depending on the clipping method of the animation. Moreover, the name of the animation needs to be calculated for the animation data that is updated at any time, and the processing capacity of the processing device becomes huge. In addition, since different images are retrieved if the extracted feature values match or are similar to the feature amount used as the basis for calculating the animation name, there is a problem of low authentication accuracy.

Therefore, an object of the present invention is to provide a collation system capable of efficiently checking browsing video data that has been made available for browsing at a site on the Internet. Another object of the present invention is to provide a method for creating an animation database that increases the number of registered animation data per unit capacity in an animation database that registers animation data as information and facilitates search. Furthermore, an object of the present invention is to provide a registration system and program that can be used in an easy-to-use method when registering animation data in an animation database, and can improve security against information leakage. In addition, an object of the present invention is to provide copyright management or a new business model using these collation systems, animation database creation methods, registration systems, and programs. These purposes are representative of those purposes that are closely related to each other, are interrelated or overlap.

The collation system of the present invention is a collation system for collation of browsing video data set to be available for browsing at a site on the network, and is characterized in that: it has a video database that registers a plurality of registered video data as information; and collates the browsing video means for collating data and registered animation data registered in the animation database, the browsing animation data having a still image displayed as a sample, the animation database having a recording medium in which each registered animation data is recorded with the A registration image group composed of a plurality of registration images generated from a plurality of frame images of reproduced animation obtained by registering the animation data, the collation unit collates the retrieval image generated from the sample still images of the browsing animation data with the An image for registration recorded in the recording medium of the video database.

Furthermore, it is preferable that the collation system according to the present invention has a function of arbitrarily or arbitrarily extracting sample still images from the browsing video data. In addition, it is preferable that a result of correlation calculation between one registration image and another registration image included in the registration image group of each registration video data is smaller than a preset threshold value. Furthermore, it is preferable that at least a binary image processing is performed while the image for registration is generated from the frame image. Here, it is preferable that at least a binary image processing is performed while the search image is being generated from a sample still image of the browsing video data, and that the collation unit performs the search image and the registration image Correlation operations to check them.

In addition, in the collation system of the present invention, image feature information of frame images corresponding to the images for registration may be recorded in the video database, and the collation unit may calculate the value of the sample still images of the browsing video data. The image characteristic information is checked with the image characteristic information recorded in the animation database. In addition, in the above collation system of the present invention, at least one keyword may be set in the registered animation data, at least one keyword data is added to the browsing animation data, and the checking unit checks the data from the browsing animation data. The keyword acquired in the added keyword data and the keyword set in the registered video data.

Furthermore, in the collation system of the present invention, it is preferable that identification data is added to the browsing animation data when the browsing animation data is registered in the animation database, and it is more preferable that the browsing animation data is changed based on the identification data. Viewing conditions of video data.

In addition, the method for making an animation database of the present invention is a method for creating an animation database in which a plurality of registered animation data are registered as information, and is characterized in that it has: a reproduction step of reproducing one of the registered animation data to obtain a reproduced animation; a generation step, A plurality of frame images of the reproduced animation is subjected to image processing including binary imaging, thereby generating a plurality of registration images; and a recording step of recording a registration image composed of the plurality of registration images on a recording medium. The image group includes a selection step between the reproducing step and the recording step. In the selecting step, a result of correlation calculation between one registration image and the other registration image included in the registration image group is smaller than a predetermined value. Selected in the way of a certain threshold.

In addition, in the animation database creation method described above, it is preferable that the selection step includes a process of performing a correlation calculation between the registration images generated from each frame image of the reproduced animation, and when the result of the correlation calculation is greater than or equal to When the preset threshold is reached, one image for registration will be deleted.

Furthermore, the method for creating a moving image database may include a registration data recording step of calculating image characteristic information for a frame image corresponding to the registration image, and recording the image characteristic information in association with the registration image.

In addition, in the method for creating a video database described above, it is preferable that at least the reproducing step and the generating step are executed in an information processing device of a registrant having one of the registered video data.

In addition, the registration system of the present invention is a registration system for registering and registering animation data in an animation database in which a plurality of registered animation data are registered as information, and is characterized in that it has a network, a server of the animation database connected to the network, and The information processing device connected to the network, wherein the information processing device regenerates one of the registered video data to obtain the reproduced video, performs at least binary image processing on a plurality of frame images of the reproduced video, and generates a plurality of frames for transmission. image, sending the plurality of images for transmission to the server of the animation database, the server of the animation database generates a plurality of images for registration from the plurality of images for transmission, and records the images generated by the plurality of images on the recording medium. A group of images for registration consisting of images for registration.

In addition, in the above-mentioned registration system, it is preferable that the image for transmission is the image for registration, and the server of the animation database records a plurality of images for transmission sent from the information processing device as a plurality of images for registration in the record. medium. Furthermore, it is preferable that the information processing device performs a correlation calculation between images based on each frame image of the reproduced animation, and deletes one of the images when the result of the correlation calculation is greater than a preset threshold value.

In addition, in the program of the present invention, in order to register the registered animation data in the animation database in which a plurality of registered animation data are registered as information, the information processing device connected to the server of the animation database through the network has the following functions: a reproduction unit, regenerating the registered animation data to obtain the regenerated animation; the image processing unit performs at least binary image processing on multiple frame images of the regenerated animation to generate an image for transmission; and the sending unit transmits the image to the The server of the animation database transmits the plurality of transmission images.

In addition, in the above program, it is preferable that the information processing device also has a function of a selection unit that performs correlation calculations between images based on each frame image of the reproduced animation, and when the result of the correlation calculation is greater than When the preset threshold is reached, one side of the image will be deleted. Here, the information processing apparatus may have a function of a threshold value input unit, and the selection unit may use the threshold value input from the input unit.

By using the checking system of the present invention, because the browsing animation data set to be browseable in the site on the network can be checked with the registration animation data registered in the animation database, the management and management of the registration animation data can be collectively executed. In addition, copyright protection or promise can be checked within the scope of registration in the animation database, which can prevent infringements caused by uploading illegal animation data and prevent registrants from exercising their rights. Other effects are described in the following embodiments.

Description of drawings

FIG. 1 is a schematic configuration diagram showing a collating system of the present invention.

FIG. 2 is a block diagram schematically showing the flow of data.

FIG. 3(A) shows a schematic data structure of an animation database, and (B) to (E) each show a typical data structure of registered animation data.

FIG. 4 is a flowchart of animation database creation processing.

FIG. 5 is a flowchart of collation processing.

Fig. 6 is a structural diagram of a verification server.

7(A) to (C) show the error rate in the animation database.

Fig. 8 shows the error rate in a collation system using still images for samples as images for retrieval.

Detailed ways

Embodiments of the present invention will be described below with reference to the drawings, but the present invention is not limited to the following examples. FIG. 1 is a schematic configuration diagram showing a collating system and a related system of the present invention, and FIG. 2 is a block diagram schematically showing data flow. In FIG. 1 , a network 11 is formed to connect the client group 10, a site 12 for providing animation data to the client group 10 is set up on the network 11, and the browsing image data 13 set to be viewable in the site 12 is checked. Checklist server 1. First, the entire collation system and its business model will be briefly described, and then the specific configuration of the collation system and the like will be described in detail.

In the collating system 1, an animation database 2 for registering a plurality of registered animation data as information, and a collating means 3 for collating the animation data and the registered animation data registered in the animation database 2 are provided. The registered animation data is provided by the registrants 5 and 6 and registered in the animation database 2 .

The registrants 5 and 6 are, for example, the operator of the site 12, or the copyright owner or creator of the registered video data. Here, as shown in FIG. 2, registration data such as the title, browsing conditions, keywords, and image feature information may be registered at the same time as the registered video data. Regarding the login data, even if it is not the registrants 5 and 6, it is possible to arbitrarily log in to the verification server or the site.

The client group 10 is composed of a plurality of clients using the site 12, is connected to the network 11, and can request the server of the site 12 to view animation data. For example, as the client group 10, a user of a video sharing site or a member of a paid content can be mentioned, and as a specific terminal device, a personal computer, a portable information terminal, a mobile phone, etc. can be mentioned.

The network 11 is an electrical communication line network connecting terminal devices of clients, and examples thereof include the Internet, WAN (Wide Area Network (Wide Area Network)), LAN (Local Area Network (Local Area Network)), and the like.

The site 12 prepares contents for providing browsing animation data to the client group 10, and in FIG. 1, contents A, B, and C having different browsing conditions are prepared. Examples of the contents of A, B, and C include completely free contents, partially viewable contents, paid contents, and the like. The site 12 may set keywords at the time of posting for the posted browsing video data. In addition, the site 12 may be configured to display at least one still image (generally called a thumbnail) as a sample of the posted browsing video data.

It is preferable to make the browsing video data 13 available for browsing in the site 12 after being checked by the checking server 1, but it may also be directly uploaded to the site 12 without going through the checking server 1 as shown in FIG. 2 . When directly uploading the browsing animation data 13 to the site 12, the checking server 1 obtains the browsing animation data from the site 12, and then checks with the registered animation data. In FIG. 1 , browsing animation data 13 is uploaded from a client group 10 to a server at a site 12 through a network 11 . The uploaded browsing animation data 15 is downloaded to the terminal device of the client through the network 11 in response to a browsing request from another client.

The browsing animation data 13 may be made available for browsing by means other than the client computer, or may be transmitted to the site 12 by means other than the network. In addition, only a link is pasted on the site 12, and the browsing animation data 13 itself is stored in the terminal device of the provider of the animation data. Here, when there is a viewing request from another client computer via the link pasted on the site 12, the download is directly performed from the terminal device of the provider to the terminal device of the other client computer.

In the collating system of the present invention, in order to perform collating processing at a higher speed, it is preferable that the browsing video data have a still image displayed as a sample. The still image displayed as a sample refers to a frame image in a reproduced animation of browsing animation data or an image reduced in size by shrinking a frame image, and is an image displayed on at least the collation server. Typically, the first frame image of the animation is reproduced, but it is not limited to this. In addition, one or more still images may be displayed as samples. Not only images displayed on the site 12 as thumbnails of the posted browsing video data, but also images that are downloaded and displayed as samples of the browsing video data when there is a browsing request from the client even if they are not displayed on the site 12 images, or still images based on arbitrary or arbitrarily extracted frame images not displayed on the client computer or network as described later. As shown in FIG. 1 , when browsing animation data 13 is uploaded from the client group 10 to the site 12 via the verification server 1, as an interface between the information processing device of the client group 10 and the information processing device of the verification server 1, as long as it is If frame images in the reproduced animation of the browsing animation data 13 can be downloaded or images reduced in size by reducing the size of the frame images, the browsing animation data 13 has still images displayed as samples.

In addition, it is preferable to add at least one keyword data (refer to FIG. 2 ) for easy checking and searching to the browsing animation data 13 . The keyword data may be arbitrarily selected by a contributor who browses the video data 13, but a keyword for specifying a field of the video data may be set in advance. In addition, it is also possible to employ a structure in which a contribution cannot be made to the site 12 without inputting a keyword. This keyword data may be used when a client computer searches for a video in the site 12, or may be used in the video database 2 of the verification server 1 for easy verification. For example, when registering animation data in the animation database 2, a keyword is recorded as the registration data, and the registered animation data can be checked from the registered animation data corresponding to the keyword data, and the registered animation can be registered as the animation database 2 in the field and keyword category. Data, collated from fields corresponding to key data.

It is preferable to give identification data to the checked-up animation data 14 checked by the checking server 1 . Regarding the identification data, it is possible to give only the history that has been verified in the verification server 1, or it may be the identification data to identify whether the browsing animation data 13 is registered in the animation database 2, and when the browsing animation data is registered in the animation database 2 At 13 o'clock, browsing conditions and part of the registration data may be included.

For example, when the browsing animation data 14 to which the identification data has been verified in the verification server 1 is assigned is downloaded to another client computer and posted again from another client computer, the verification server 1 recognizes that it has been verified. With such identification data, it is possible to determine whether or not to collate the browsing video data 14 to which the identification data is added. In addition, if the identification data includes information on whether or not the animation data is registered in the animation database 2, the browsing conditions can be changed according to the presence or absence of registration, and the animation data on the network can be managed. Furthermore, if the identification data further includes information such as the title of the animation and viewing conditions, more detailed setting and management of viewing conditions can be performed.

Browsing conditions include, for example, browsing animation data 13 that is posted without being browsed and deleted, paid content, billing based on the number of times viewed, only partially viewable, or invalid content.

In FIG. 1, the browsing animation data 13 is checked with the animation database 2 of the checking server 1, the identification data of A~C are given according to the checking result, and the identification data of A~C are uploaded to A, B, and C prepared on the site 12. of each content.

It is also possible to provide identification data only to the browsing video data registered in the video database 2 among the browsing video data. In addition, when the browsing video data corresponding to the collation result is processed in the collation server 1, identification data may not be provided to the browsing video data. For example, if the registered animation data registered in the animation database 2 cannot be uploaded unless all of them are browsed, the browsing animation data that matches the registered animation data registered in the animation database may be deleted in the verification server 1, No identification data may be added to the browsing video data uploaded to the site 12 . In other words, uploading itself to the site 12 means that the browsing animation data is not registered in the animation database. In addition, if it is only to conduct a simple survey or manage the number of times of posting video data, it is only necessary to provide such information to the registrant in the verification server 1, and identification data may not be provided to the browsing video data 13 .

Furthermore, when the browsing video data 13 is registered in the video database 2, the checking server 1 or the site 12 can provide information related to the video data to the registrant who has registered the video data in the video database 2 (see FIG. 2 ). The information related to video data includes, for example, the number of times registered video data has been posted, the number of times it has been viewed (downloaded), information on a client computer that has posted video data, and the like.

In addition, the check server 1 may transmit a warning message notifying copyright infringement of the posted animation data to a client computer that has posted animation data registered in the animation database 2 as the browsing animation data 13 (see FIG. 2 ). By sending such a warning message, the client group 10's awareness of copyright can be raised.

When using the collation system as above, the registrants 5 and 6 can manage the registered animation data on the network 11 collectively or register the copyright by registering the animation data they want to manage in the animation database 2 as the registered animation data. protection or promise. In addition, the site 12 can check the browsed animation data within the range registered in the animation database 2, thereby avoiding copyright infringement caused by uploading illegal animation data and preventing the registrant from exercising rights.

That is, the collation server 1 can provide a service of registering animation data to be managed in the animation database 2 to the registrants 5 and 6, and can also provide a service of substituting registered part of the animation data for management. Furthermore, the verification server 1 may provide the site 12 with a service of checking the browsing video data within the range registered in the video database 2, and may also provide a service of substituting a part of the browsing video data for management. The check server 1 can also be provided by the registrant 5 , 6 or the site 12 itself.

Then, the collation server 1 is explained. The verification server 1 is provided with a recording medium and an information processing device, the recording medium records the animation database 2 registered with a plurality of registered animation data as information, and the information processing device has a check animation data and an information processing device registered in the animation database 2. At least a part of the function of the collating means 3 of the registered video data.

One of the features of the collation system based on the collation server 1 of the present invention is that the final collation between browsing video data and registered video data is performed by comparing images (including electronic image data). That is, although the collation system of the present invention does not exclude the conventional collation by keywords or feature values, it is necessary to perform a search for a registration image generated from a reproduced animation of the registered animation data and a retrieved image generated from a reproduced animation of the browsing animation data. Direct checks with images. Since the images are directly collated with each other, the collation accuracy can be improved. In addition, as the registered video data, by recording a registration image group composed of registration images corresponding to the main frame images of the reproduced video of the registered video data, even if the search image is any scene of the registered video data, it can be reliably performed. search.

Therefore, even when checking is performed using only the search image generated from the sample still image of the registered video data, it can be determined whether the browsing video data and the registered video data match or not by checking with the registration image group of the registered video data. Is it similar. The still image itself for the sample has an extremely small capacity compared with the downloaded browsing video data, so the download time can be shortened, and high-speed verification can be performed. Therefore, it becomes valid collation data even for a plurality of browsing animation data uploaded to the site. For example, even if the capacity of the browsing movie data is several MB (megabytes), since the thumbnail itself is image data of about 128 pixels (horizontal) x 96 pixels (vertical), it is possible to absolutely shorten the download time and computation time. quantity.

In this manner, in a collation system using thumbnails, there is a possibility that a malicious client may interfere with collation by performing some kind of image processing on the frame images used as thumbnails. Therefore, it is preferable that the collating system also has a function of arbitrarily or arbitrarily extracting sample still images from the browsing video data. In order to realize such a function, the information processing device of the uploading client, the information processing device of the check server, and the information processing device of the server of the site, etc. have a reproduction unit for reproducing the browsing animation data to obtain the reproduction animation, and from the regeneration animation The program may be a program that functions as an image generating unit that extracts an arbitrary frame image or extracts a frame image at will, and generates a still image of a sample from the extracted frame image. Here, it is preferable to generate a plurality of sample still images from a plurality of temporally spaced frame images. In addition, in order to improve the reliability of the collation system, it is preferable to use a program that cannot be operated on the client computer for the extraction position of the frame image. In addition, it is also effective to adopt a configuration in which which frame image is extracted cannot be viewed on the client computer or the network.

Then, in the moving image database 2, as each registered moving image data, a registration image group consisting of a plurality of registration images generated from a plurality of frame images of a reproduced moving image obtained by reproducing the registered moving image data is recorded. That is, as shown in FIG. 3(A), in the animation database 2, a plurality of registration image groups 16 corresponding to the number of registered animation data are recorded, and in each registration image group 16, a plurality of registration image groups 16 are respectively included. 17.

As a plurality of frame images, all the frame images of the playback movie are registered as they are, and the maximum amount of information of the registered movie data is recorded. As a result, the calculation time becomes longer and a huge storage capacity is required. Animation currently used in Japan is 29.97fps (frame per second) in the NTSC standard for TV broadcasting and 24fps for movies, so it consists of about 30 frames and 24 frames of still images in one second. In addition, since each frame image has graphic information of an arrangement of pixels corresponding to the resolution and color information for each pixel, when all frame images are registered as a registration image group, the total data capacity become enormous. For example, a full-color (24-bit) image with a resolution of 600×480 is simply calculated as a data capacity of 27MB, and when it is a 90-minute movie (24fps), there are 129,600 frame images, so even if calculated by one animation data It's not realistic by any means.

Therefore, it is preferable to generate a registration image by performing at least binary image processing on the frame image of the reproduced video. The so-called binary image processing refers to the processing of converting into an image displaying each pixel in two colors of white and black (0 and 1). Generally, each pixel of animation data has at least gradation information (8 bits), and most Since it has the color information (24 bits) of the color, the data capacity of the image for registration can be significantly reduced by binary image processing. In addition, when the binary registration image subjected to binary imaging processing and the binary search image similarly subjected to binary imaging processing are performed to perform correlation calculations for comparison, the calculation speed is faster than that of the full-color image. When performing correlation calculations between images or between grayscale images, the calculation speed is fast, and the matching accuracy may be improved in some cases, and binary image processing is extremely useful. Furthermore, since the amount of computation of the image subjected to the binary image processing is small, it is also possible to manufacture a dedicated computation device using integrated circuits such as LSI as hardware.

As methods of binary image processing, they can be roughly classified into two types: a fixed threshold method and a region specifying method. The former is a method of determining the threshold value for binarization using a certain value of the gray scale, and the latter is a method of determining the ratio of binary values of the entire image in advance, and then determining the threshold value of binarization from the histogram of density values. When the threshold of binarization is T, the element of each pixel of the original image is f(i, j), and the element of each pixel of the binary image is g(i, j), the following formula (1) Defined that way.

[mathematical formula 1]

$\mathrm{<span\; class="notranslate">\; g</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\left\{\begin{array}{cc}\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&Greater\; Equal;</span>\mathrm{<span\; class="notranslate">\; T</span>}\\ \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; T</span>}\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

When the frame image is a color image, first remove the color information of the color, change it to grayscale with only level information, and then perform binarization. In addition, for frame images or binary images, other image processing can also be performed. For example, image processing for changing the resolution, edge emphasis processing, etc. may be performed before or after the binary image processing.

In addition, the binary image processing is basically an irreversible conversion that only deletes information, unlike data format conversion processing, compression processing, and encryption processing. That is, because of data format conversion processing, compression processing, and encryption processing, in order to restore the original frame image, the information that can be inversely converted is itself or another file. However, the binary image processing discards the color information and gradation information of the original frame image, and the binary image itself cannot restore the original frame image.

The inability to restore the original frame image may be considered a disadvantage at first glance, but it is actually effective in preventing information leakage of the original animation data. That is, in order to register the original animation data from the registrant to the animation database, when it is provided to the verification server as registered animation data, the original animation data may flow out due to an accident in distribution, an accident in network transmission, or artificial information leakage. dangerous. For example, the loss caused by the release of public movies or unpublished life images is huge. However, if the registrant's information processing device performs at least binary image processing on the frame image of the registered video data, and then provides the binary image to the verification server, even if the binary image flows out, it cannot be restored by itself. It is the original animation data, so the outflow of original animation data can be avoided. A registration system and program for registering animation data in the animation database using this effect will be described later.

In addition, in order to reduce the number of images for registration, it is preferable to delete similar images and not to register them, but to select and register them to some extent. For this purpose, a correlation operation is performed between each frame image of the registered animation data, or between images (including binary images) obtained by image processing of each frame image, and the result is greater than a preset threshold When , one of the images is excluded, and if the result is smaller than the threshold value, at least one of the images is registered. The threshold is a criterion of similarity for selection. If the threshold is high, the collation accuracy generally increases, but the number of images for registration also increases; if the threshold is low, the collation accuracy decreases, but the number of images for registration also decreases, so it is preferable to The optimum threshold value is determined by experiments or the like. In addition, if the registrant can change the threshold, the threshold can be set higher for important video data, and the threshold can be set lower for video data of low importance. One of the images for registration does not need to be correlated with all the other images for registration, and it is only necessary that the result of the correlation calculation with at least one other image for registration is smaller than a threshold value.

As a correlation calculation, for example, the degree of similarity can be calculated by obtaining the inner product, the Euclidean distance, or the correlation coefficient between registration images. Regarding the elements of each pixel of the registration images F and G of N×N pixels, when expressed as a one-dimensional column vector f and a column vector g, the inner product can be calculated by formula (2), and the inner product can be calculated by formula (3) To calculate the Euclidean distance, the correlation coefficient can be calculated by mathematical formula (4). Calculation of the degree of similarity is not limited to these methods.

[mathematical formula 2]

$<span\; class="notranslate">\; <</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{\mathrm{<span\; class="notranslate">\; f</span>}}<span\; class="notranslate">\; ,</span>\stackrel{<span\; class="notranslate">\; \&Right\; Arrow;</span>}{\mathrm{<span\; class="notranslate">\; g</span>}}<span\; class="notranslate">\; ></span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}{\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

[mathematical formula 3]

${\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; e</span>}}<span\; class="notranslate">\; =</span>\sqrt{{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

[mathematical formula 4]

$\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}{\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}}{\sqrt{{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="2"\; label="f\; k"\; filenames="HPA00001049603700031.png"\; state="\{\{state\}\}">f</figure-callout></span><figure-callout\; id="2"\; label="f\; k"\; filenames="HPA00001049603700031.png"\; state="\{\{state\}\}">\; </figure-callout>}}_{\mathrm{<span\; class="notranslate">\; </span>}}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}\sqrt{{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}^{\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{{\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

3(B) to (E) schematically show the data structure of each registered video data described above. FIG. 3(B) uses all the frame images of the reproduced animation as the registration image group 16, and each registration image 17 is an image of each frame, and has level information or color information. FIG. 3(C) performs binary image processing on FIG. 3(B). Since each registration image 17 becomes a binary image, the level information or color information is removed, and the data capacity of the registration image group 16 can be reduced. FIG. 3(D) is selected with respect to FIG. 3(B), and the selected frame images of the respective registration images 17 can reduce the data capacity of the registration image group 16 corresponding to the capacity of the deleted frame images. FIG. 3(E) performs binary image processing and selection of both for FIG. 3(B). Binary image processing can be performed first, and then the binary image can be selected (Figure 3(B)→(C)→(E)), or the binary image processing can be performed after the frame image is selected (Figure 3(B) →(D)→(E)). In FIG. 3(E), each registration image 17 is a selected binary image, and the data capacity can be significantly reduced compared to FIG. 3(B).

The manufacturing method of such an animation database only needs to have the following steps: a regenerating step of regenerating one of the registered animation data to obtain a regenerated animation; a generating step of at least performing binary image processing on a plurality of frame images of the regenerated animation, Generate a plurality of images for registration; Recording step, recording on the recording medium a registration image group consisting of a plurality of registration images; Selection step, between the reproduction step and the recording step, registering with one of the registration images contained in the recording medium The selection is made so that the result of the correlation calculation between the image and another image for registration is smaller than a preset threshold value. As described above, any one of the generation step and the selection step including binary image processing may be performed first, and other image processing or steps may be performed between or before and after these steps. However, in order to speed up the calculation speed and improve the accuracy of the correlation calculation processing in the selection step, it is preferable to perform the correlation calculation in the selection step after performing the binary image processing.

FIG. 4 is an example of a flowchart of animation database creation processing. First, registered video data is acquired, and the registered video data is reproduced (S41: reproduction step). Generate all frame image groups M _i (i=1, 2, 3, ... N, N is the number of frames) from the regenerated animation, for all frame image groups M _i , perform binary image processing, and convert to binary images Group B _i (i=1, 2, 3, . . . N, where N is the number of frames) (S42 to S43: generating step). Then, the reference image B _p used as a comparison reference is substituted into _Bi (initial value i=1), and the correlation calculation between the reference image B _p and the binary image B _i (i=i+1) to be compared is performed to obtain a correlation value. V (S44-S46: selection step). When the correlation value V is above the threshold (No of S47: selection step), it is judged that the reference image B _p is similar to the binary image B i, the binary image _{B i is} deleted and not registered in the animation database, and the reference image B _p is performed. Correlation calculation with the next binary image B _i (i=i+1) (S45-S46). On the other hand, when the correlation value V is lower than the threshold (Yes in S47: selection step), it is judged that the reference image B _p is not similar to the binary image B _i , and the reference image B _p is registered in the animation database (initial value is binary value image B _i ) (S48: recording step), the binary image B _i that is not similar is used as the next reference image B _p and the next binary image B _i (i=i+1) is correlated (No of S49 , S44～S46). When the binary image B _i is the binary image B _N of the last frame in S49, the registration of the frame image B _N then ends (S50). Finally, a plurality of reference images registered as dissimilarities are recorded as a registration image group B _m (m=1, 2, 3, . . . M, where M is the number of registration images).

In the algorithm of the flow chart shown in FIG. 4 , the initial reference image and the next reference image are correlated. However, correlation calculations with other reference images may not be performed and similar images may be registered. Regarding this point, if for the initial reference image, perform correlation calculations on all binary images other than it in order, delete the binary images whose correlation value is greater than the threshold, do not log in the animation database, and replace the binary images with larger thresholds If one of the images is used as the next reference image, and the other binary images are sequentially correlated, the number of registration images can be further reduced.

In addition, a registration data recording step may be included to record registration data such as titles corresponding to the registration image group, browsing conditions, keywords, and image feature information on frame images corresponding to the registration images. As image feature information, when the image for registration is a binary image or grayscale, if there are features in the distribution of specific colors or areas of the original frame image, these features are recorded as information, even for the image for retrieval The image feature information is also calculated for the corresponding frame image, and the image feature information can be used as a secondary check.

The creation of the video database may be performed only by the verification server, or a part of the steps may be performed in the registrant's information processing device. In order to prevent the above-mentioned information leakage, it is preferable to perform the reproduction step of obtaining the reproduced moving image and the generation step of binary image processing for the frame image by the registrant's information processing device. The image for registration or the image for transmission generated by the information processing device of the registrant may be sent to the verification server via a network, or may be sent to the verification server offline.

In addition, when describing an example of a login system using a network, the login system includes a network, an information processing device connected to the network, and a video database server. Then, the information processing device reproduces one of the registered video data to obtain the reproduced video, and performs at least binary image processing on a plurality of frame images of the reproduced video to generate a plurality of transmission images. As the image for transmission, at least binary image processing may be performed on the frame image, and other image processing may also be performed. In the server of the animation database, it may be recorded as the image for registration as it is, or it may be This is the state recorded in the video database server after further processing such as image processing. Then, the information processing device transmits a plurality of transmission images to the server of the animation database via the network. Since multiple images for transmission are converted into binary images by binary image processing, even if they are leaked on the network, they will be irreversibly changed to the state of reproduced animation, which can prevent information leakage of reproduced animation. The video database server generates a plurality of images for registration from the received images for transmission, and records a group of images for registration consisting of the images for registration on a recording medium. When the received multiple transmission images can be used as the registration images without change, the video database server records the multiple transmission images without change on the recording medium as the multiple registration images.

In addition, the information processing device may perform a correlation calculation between images based on each frame image of the reproduced animation, and may delete one of the images from the transmission image when the result of the correlation calculation is greater than a preset threshold value. This is not performed by the server of the animation database in the selection step, but is performed by an information processing device connected via a network, so that the total data capacity of images for transmission can be reduced. An image based on a frame image includes a frame image itself, an image subjected to image processing other than binary image processing, and an image subjected to binary image processing.

In order to utilize such a registration system, an information processing device is equipped with a reproducing unit for reproducing registered video data to obtain a regenerated video, and performs at least binary image processing on a plurality of frame images of the reproduced video to generate an image for transmission. The image processing unit and the program of the function of the sending unit that sends a plurality of sending images to the server of the animation database via the network may be sufficient. In addition, the information processing device may have a function of selecting means for performing a correlation calculation between images based on each frame image of the reproduced animation, and deleting one of the images when the result of the correlation calculation is greater than a preset threshold value. Here, in order for the registrant to select the threshold value, the information processing device may have a threshold value input unit function.

However, a large-capacity hard disk, a holographic recording medium, or the like can be used as the recording medium of the moving image database. When using a large-capacity hard disk in the video database and collating images by arithmetic processing, an information processing device can be used as a collating unit. On the other hand, when using a hologram recording medium, an optical recording unit and a collation unit are required. The above-mentioned collation system, animation database, registration system, etc. can be used in any collation unit.

Next, the operation of the collating means for collating will be described. As described above, in the collating system of the present invention, it is preferable to perform collating using a search image generated from a sample still image of browsing video data. Fig. 5 is a flowchart of collation processing when an information processing device performs collation using a search image generated from a sample still image of browsed video data in a video database in which a group of registration images subjected to binary image processing is recorded picture.

First, the information processing device of the collation server acquires a still image for a sample of browsing video data (S51). Next, binary imaging processing is performed on the still image, and converted into a binary search image S (S52). Then, for the binary retrieval image S and each registration image of the registration image group B _m (m=1, 2, 3, . . . M, M being the number of registration images) recorded in the animation database, Correlation operations are performed sequentially to obtain the maximum correlation value V _max (S53). The maximum correlation value V _max is compared with the verification threshold of the video data, and when the maximum correlation value V _max is greater than the verification threshold (Yes in S54: verification step), the browsing video data is recognized as the registration video data of the image group for registration (S55). In addition, when the maximum correlation value V _max is equal to or less than the verification threshold (No: verification step of S54), it is recognized that the browsing video data is not the registration video data of the registration image group, and then it is determined whether there is a next registration image group (S56). . When the next image group for registration exists (No in S56), the correlation calculation is performed again in order for the image S for retrieval and each registration image in another image group for registration to obtain the maximum correlation value V _max and compare it with the verification threshold. (S53～S54). Next, when there is no image group for next registration (Yes in S56), it is recognized that the browsing video data is not registered in the video database (S57).

In the processing (S52) of converting the still image into a binary image S by converting it into a binary image (S52), other image processing can be performed. For example, image processing for changing the resolution or edge enhancement processing may be performed before or after the binary image processing. In addition, when selecting an image group for registration to be compared with the search image S, keyword data of browsing video data or image feature information of sample still images may be obtained, and the registered data may be referred to to preferentially select the registered keyword data. image group for registration, or preferentially select a group of registration images including images for registration whose image feature information matches.

Furthermore, the information processing device of the collation server may add a collation result as identification data to the browsing video data that has been collated, and then upload it to the site.

The collation server 1 using the holographic recording medium will be briefly described below.

FIG. 6 shows a specific configuration of a verification server 1 using a holographic recording medium, and includes a holographic recording medium 20 , an information processing device 30 , and an optical system 40 . An image database 2 and a collation unit 3 are provided in the collation server 1 , the holographic recording medium 20 constitutes a part of the image database 2 , and the information processing device 30 and the optical system 40 constitute the collation unit 3 .

In FIG. 6 , the hologram recording medium 20 is a reflective recording medium in which a thick hologram recording layer 21 made of a photosensitive material is sandwiched between a surface protection layer 22 and a reflective layer 23 . In the hologram recording layer 21 of the hologram recording medium 20, image data is recorded by interference fringes 24 of information light generated based on the registered image data and reference mark light generated based on the reference mark. When an original disk-shaped medium is used as the hologram recording medium 20, since the collation can be performed while rotating, the collation speed can be increased. In addition, it is preferable that the hologram recording medium 20 has an address layer for specifying the positions of the interference fringes 24 . For example, as address information, it is possible to form pits in a concavo-convex shape provided on the surface of the reflective layer 23, and the reflective layer 23 can be used as an address layer. When a glass substrate is used as the surface protection layer 22 of the hologram recording medium 20, shrinkage or the like due to temperature changes or the like can be suppressed. A metal material such as aluminum can be used as the reflective layer 23 .

The information processing device 30 is connected to a network, and performs various information processing performed by the collation system. Examples include reproduction of browsing image data and creation of search images, recording and reproduction of registration data, adding identification data to browsing video data, determining a collation order based on keyword data, transmission of information related to video data, and warnings. Message sending, etc.

The optical system 40 is capable of recording interference fringes on the holographic recording medium 20 and collating the registration image data and review image data recorded on the holographic recording medium 20 . The optical system 40 has a holographic laser 41, a mirror 42, a spatial light modulator 43, a polarizing beam splitter 44, a first relay lens 45, a mirror 46, a second relay lens 47, a beam splitter 48, a quarter Wave plate 49 , objective lens 50 , aperture 51 , reference photodetector 52 , address laser 60 , beam splitter 61 , mirror 62 and address photodetector 63 .

The holographic laser 41 serves as a light source for information light and reference light when recording interference fringes, and as a light source for search light when checking. For example, a short-wavelength high-output laser such as a blue laser or a green laser is preferable. The spatial light modulator 43 has a plurality of pixels, and can spatially modulate light by changing the property of light for each pixel, for example, a liquid crystal display device or a DMD (Digital Micromirror Device) can be used. In FIG. 6 , a DMD is used as the spatial light modulator 32 . The polarizing beam splitter 44 transmits one of the orthogonal polarization directions and reflects the other, transmits the information light, reference light, and search light toward the recording medium 20, and reflects the reference light reproduced by the recording medium toward the reference photodetector 52. Light. The first and second relay lenses 45, 47 image the image displayed on the spatial light modulator 43 on the focal plane of the objective lens. The beam splitter 48 directs the light from the address laser 60 toward the recording medium 20 . The quarter-wavelength plate 49 is used to convert linearly polarized light into circularly polarized light, and the linearly polarized light can be rotated by 90 degrees through secondary transmission. Through this quarter-wavelength plate 49 , the reference light passes through the polarizing beam splitter 44 during irradiation, and is reflected through the polarizing beam splitter 44 during reproduction. The objective lens 50 is used to perform Fourier transform on the image displayed on the spatial light modulator 43 , and then irradiates on the hologram recording layer 21 of the recording medium 20 . The diaphragm 51 has an opening for blocking the search light reflected by the recording medium 20 and passing only the reproduced reference light to the reference light detector 52 . The reference light detector 52 is used to detect the reproduced reference light, and it is particularly preferable to detect the light intensity of the reference light. For example needle photodiodes, CMOS sensors or CCD sensors can be used.

The light irradiated from the holographic laser 41 is reflected by the mirror 42 and spatially modulated by the spatial light modulator 43 for displaying an image for registration or an image for search, thereby generating information light or search light. The information light or search light passes through the polarizing beam splitter 44, is relayed by the first and second relay lenses 45, 47, is reflected by the mirror 46 on the way, and passes through the beam splitter 48 and the quarter-wavelength plate 49 , perform Fourier transform by the objective lens 50 , and irradiate onto the hologram recording layer 21 of the recording medium 20 . The information light interferes with the reference light in the hologram recording layer 21 to record interference fringes. When the search light has a high correlation value with the information light on which the interference fringes recorded on the hologram recording layer 21 are high, it is diffracted by the interference fringes to reproduce the reference light. The reproduced reference light is reflected by the reflective layer 24, exits the recording medium 20, and passes through the objective lens 50, the quarter-wavelength plate 49, the beam splitter 48, the first and the second relay lens 45 in the direction opposite to the irradiation time. , 47 , mirror 46 , reflected by the polarizing beam splitter 44 , and incident to the reference photodetector 52 through the aperture 51 .

In addition, the address laser 60, the beam splitter 61, the mirror 62, and the address photodetector 63, when an address layer is provided on the recording medium 20, determine the irradiation position from the address layer, and the light irradiated from the address laser 60 transmits It is reflected by the mirror 62 through the beam splitter 61 and reflected by the beam splitter 48 , transmitted through the quarter wave plate 49 , and irradiated on the address layer of the recording medium 20 through the objective lens 50 . Reflected light from the recording medium 20 , on the contrary, passes through the optical system, is reflected by the beam splitter 61 , and is detected by the address photodetector 63 . As the address laser 60, it is preferable to use a laser with a relatively long wavelength such as red light.

In the verification server 1 using such a holographic recording medium, when recording the image for registration in the moving image database, the image for registration and the image for reference light are displayed on the spatial light modulator 43, and the image for registration and the image for reference light are passed through. Spatially modulated information light and reference light are generated. Then, the information light and the reference light are Fourier transformed by the objective lens 50, and then irradiated onto the hologram recording layer 21 of the recording medium 20, whereby the Fourier transformed information light and the reference light interfere in the hologram recording layer 21, Interference fringes corresponding to the image for registration are recorded. Then, when checking the video data for viewing, the search image is displayed on the spatial light modulator 43, and spatially modulated search light is generated from the search image. After the search light is Fourier transformed by the objective lens 50, it is irradiated on the interference fringes of the hologram recording layer 21 of the recording medium 20. , the search light is diffracted by interference fringes, and the reference light is reproduced. That is, the correlation value can be obtained by detecting the intensity of the reproduced reference light. When the image for registration or the image for search is a binary image that has been converted into a binary image, it is desirable that it is easily associated with each pixel of the spatial light modulator 43 .

In the above description, since a collinear system device in which information light and reference light are coaxially interfered is used, high-speed collation can be performed. In the above description, a reflection-type hologram recording medium has been described, but a transmission-type hologram recording medium that transmits and displays a reference mark reproduced by optical correlation calculation is also possible. Also, instead of the collinear method, a device of the two-beam interference method that separates the optical path of the information light from the optical path of the reference mark and intersects at a fixed angle on the recording medium can also be realized. For example, the light from the light source 41 can be split into two lights by a beam splitter, one light can be modulated by a spatial light modulator to generate information light, and the other light can be deformed to generate a reference mark light for recording Irradiation is carried out in a crosswise manner on the medium.

(Example 1)

As registered animation data, 30 pieces of animation data of 10 seconds and 30 fps are prepared, and all frame images (300 frames) are subjected to image processing such as binary image processing to generate a binary image group Bi. Then, the correlation value of autocorrelation (correlation calculation between the same image) is normalized to 100 (a.u.: arbitrary unit), and the threshold is set to 70 (a.u.) and 60 (a.u.), and the selection process is performed to create animation database. When the threshold value is 70 (a.u.), for 9000 binary images of 300 frames × 30, the number of registration images can be reduced to 2504. When the threshold value is 60 (a.u.), compared to 300 frames With 9000 binary images × 30, the number of registration images can be reduced to 1580. Table 1 below shows the number of cuts for each animation data.

[Table 1]

Then, the error rate is obtained for each of the animation database in which all binary images are registered, the animation database in which the selection procedure of the threshold value 70 is performed, and the animation database in which the selection procedure of the threshold value 60 is performed. 7 (A) to (C) each graph shows the error rate in each animation database, the vertical axis is the error rate, and the horizontal axis is the threshold value related to the normalized correlation value. The error rate consists of two error curves, one is the registered image rejection rate (FRR: False Rejection Rate) when a registered image is mistaken for an unregistered image, and the other is when a different image is mistaken for a registered image Other image acceptance rate (FAR: False Acceptance Rate). The registered image rejection rate FRR and the other image acceptance rate FAR are respectively extracted based on the true/false rate of the collation result when the threshold value is changed. Figure 7 (A) to (C) is to extract 1 frame of image from the registered 30 animation data to make 30 images for retrieval, and check them with the registration images of each animation database respectively. When the obtained correlation value is greater than When the threshold value is judged to be consistent with the registration image, the registration image rejection rate FRR and other image acceptance rates FAR.

The value of the intersection of the login image rejection rate FRR and other image acceptance rates FAR is the threshold at which both the login image rejection rate FRR and other image acceptance rates FAR become smaller. The error rate at this time is called EER (Equal Error Rate). When the EER has a certain range, this range is called a threshold region. The smaller the EER, the better. If the EER is 0%, then the login image rejection rate FRR and other image acceptance rates FAR are both 0%, theoretically no error occurs. In addition, the width of the threshold area indicates the reliability of the verification, and the wider the threshold area, the higher the reliability of the verification. In order to also perform similar checks on images, the threshold value is deliberately set to a value smaller than the threshold area, so as to increase the acceptance rate of other images by FAR.

Fig. 7 (A) is the situation of the animation database in which all binary images are registered, (B) is the situation of the animation database in which the selection step of the threshold value 70 is performed, and (C) is the animation database in which the selection procedure of the threshold value 60 is performed situation. The EER is 0% in Fig. 7(A) and (B), and an extremely low value of 3.3% is obtained in Fig. 7(C).

[Example 2]

Next, a collation accuracy evaluation experiment of a collation system using a search image generated from a sample still image of browsing video data was performed. As the animation database, as the registered animation data, prepare 30 animation data of 10 seconds and 30 fps, and perform image processing such as binary image processing on all frame images (300 frames) to generate 9000 binary images B _i , using An animation database with all binary images registered. The size of the still image for the sample used in the search is 128×96 pixels, and an image obtained by performing image processing such as binary image processing on one sample still image for each registered video data, a total of 30, is used as the image for the search use. FIG. 8 shows the error rate in this collation accuracy evaluation experiment, the vertical axis is the error rate, and the horizontal axis is the threshold value related to the normalized correlation value. Regarding the error rate, it is the same as in FIG. 7 . This result proves that high-accuracy collation of ERR0% can be performed.

Claims (19)

1. a check system is checked the reading animation data that is changed to the state of can reading in the website on network, it is characterized in that,

Have the cartoon databank of a plurality of login animation datas as information registration; And check described reading animation data and the login animation data in described cartoon databank, logined check the unit,

Described reading animation data has as sample to be used and the rest image of demonstration,

Described cartoon databank has recording medium, logins the login image sets that cartoon data recording is made up of with image a plurality of logins of a plurality of two field pictures generations of the regeneration animation that is obtained by this login animation data of regeneration as each,

The login image that writes down in the recording medium of retrieval with image and described cartoon databank that is generated with rest image by the sample of described reading animation data is checked in the described unit of checking.

2. check system according to claim 1 is characterized in that,

Has the function of from described reading animation data, extracting the rest image of sample arbitrarily or optionally.

3. check system according to claim 1 and 2 is characterized in that,

About the described login image sets of respectively logining animation data, this login uses image and another result who logins the related operation of using image less than pre-set threshold with a login that comprises in the image sets.

4. according to any one described check system of claim 1 to 3, it is characterized in that,

Generate by two field picture described login with image during, carry out the bianry image processing at least.

5. check system according to claim 4 is characterized in that,

The sample by described reading animation data generate with rest image described retrieval with image during, carry out the bianry image processing at least,

The described unit of checking carries out related operation with image and described login with image to described retrieval and comes they are checked.

6. according to any one described check system of claim 1 to 5, it is characterized in that,

In described cartoon databank, write down and the image feature information of each login with the corresponding two field picture of image,

The described unit of checking calculates the image feature information of the sample of described reading animation data with rest image, checks with the image feature information that writes down in described cartoon databank.

7. according to any one described check system of claim 1 to 6, it is characterized in that,

In described login animation data, set at least one key word,

Described reading animation data has added at least one key data,

Key word of obtaining and the key word of setting are checked in the described unit of checking in described login animation data from the additional key data of described reading animation data.

8. according to any one described check system of claim 1 to 7, it is characterized in that,

When in described cartoon databank, having logined described reading animation data, to described reading animation data additional identification data.

9. check system according to claim 8 is characterized in that,

Change the reading condition of described reading animation data according to described recognition data.

10. the method for making of a cartoon databank, described cartoon databank as information registration a plurality of login animation datas, it is characterized in that,

Have:

Step for regeneration, one of regeneration login animation data obtains the animation of regenerating;

Generate operation, comprise the Flame Image Process of bianry imageization, generate a plurality of login images thus for a plurality of two field pictures of described regeneration animation; And

The record operation, the login image sets that record is made up of with image described a plurality of logins on recording medium,

Possess the selection operation between described step for regeneration and described record operation, this selection operation is selected less than the mode of pre-set threshold with the result of another login with the related operation of image with image with a login that comprises in the image sets with described login.

11. the method for making of cartoon databank according to claim 10 is characterized in that,

Described selection operation comprises following processing: carrying out related operation in each login that each two field picture by described regeneration animation generates each other with image, in the result of related operation during greater than pre-set threshold, a side login image-erasing.

12. the method for making according to claim 10 or 11 described cartoon databanks is characterized in that,

Comprise logon data record operation, about with described login with the corresponding two field picture of image, the computed image characteristic information, with this login with the related ground of image document image characteristic information.

13. the method for making according to any one described cartoon databank of claim 10 to 12 is characterized in that,

In the signal conditioning package of the registrant with one of described login animation data, carry out described step for regeneration and described generation operation at least.

14. a login system, be used for as information registration the cartoon databank login login animation data of a plurality of login animation datas, it is characterized in that,

The server of the cartoon databank that have network, is connected and the signal conditioning package that is connected with described network with described network,

One of described signal conditioning package regeneration login animation data is obtained the regeneration animation, for a plurality of two field pictures of described regeneration animation, at least carry out the bianry image processing, generate a plurality of transmission images, described a plurality of transmissions are sent to the server of described cartoon databank with image

The server of described cartoon databank generates a plurality of login images by described a plurality of transmissions with image, the login image sets that record is made up of with image described a plurality of logins on recording medium.

15. login system according to claim 14 is characterized in that,

Described transmission image is described login image, and the server of described cartoon databank uses a plurality of transmissions that send from described signal conditioning package image recording with image recording medium as a plurality of logins.

16. according to claim 14 or 15 described login systems, it is characterized in that,

Described signal conditioning package carries out related operation each other at the image based on each two field picture of described regeneration animation, in the result of related operation during greater than pre-set threshold, a side image-erasing.

17. a program, for as information registration login login animation data in the cartoon databank of a plurality of login animation datas, make the signal conditioning package that is connected with the server of described cartoon databank by network have following function:

Regeneration unit, the described login animation data of regenerating obtains the regeneration animation;

Graphics processing unit for a plurality of two field pictures of described regeneration animation, carries out the bianry image processing at least, generates to send to use image; And

Transmitting element sends described a plurality of transmission image via described network to the server of described cartoon databank.

18. program according to claim 17 is characterized in that,

Also make described signal conditioning package have the function of selected cell, described selected cell carries out related operation each other at the image based on each two field picture of described regeneration animation, in the result of related operation during greater than pre-set threshold, a side image-erasing.

19. program according to claim 18 is characterized in that,

Also make described signal conditioning package have the function of the input block of threshold value, in described selected cell, use from the threshold value of described input block input.

Images